It has been shown in the literature that depression has a significant negative correlation with employment outcomes as measured by labor force participation, earnings, work attendance and job performance. I expand the understanding of this relationship by exploring the effect of depression on employment choices as well as treatment choices over time rather than simply examining correlations at a point in time. Other health related outcomes and the relationship between choices and mental health will be examined. My analysis follows initially depressed individuals for nine months and examines the dynamic relationship between health status and function, treatment decisions and employment outcomes. I consider a dynamic model of individual decisions over time where lagged endogenous behavior is allowed to influence current behavior or health outcomes. Results indicate that depression does have a significant effect on labor productivity. Individuals who were the most depressed at the baseline interview saw the largest improvements in productivity following treatment. However, the estimates imply that depression is not a significant determinant of the worker's attendance at work.